CN113017627B - Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion - Google Patents
Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion Download PDFInfo
- Publication number
- CN113017627B CN113017627B CN202011633891.6A CN202011633891A CN113017627B CN 113017627 B CN113017627 B CN 113017627B CN 202011633891 A CN202011633891 A CN 202011633891A CN 113017627 B CN113017627 B CN 113017627B
- Authority
- CN
- China
- Prior art keywords
- phase
- depression
- tested
- bipolar disorder
- electrode
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
- A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
Abstract
The invention discloses a depression and bipolar disorder brain network analysis method based on dual-channel phase synchronization feature fusion. Compared with the prior art, the method provided by the invention has the following advantages: the three synchronous features are fused, so that more effective information can be obtained to detect the weak interaction relation among signals, and the discovery of whether the electrode signals of the brain area are in a synchronous state is facilitated, so that the diseased brain area can be effectively identified. Experiments show that compared with a healthy control group, the method provided by the invention can intuitively show that the difference between the frontal lobe and the apical lobe of patients with depression and bipolar disorder is large.
Description
Technical Field
The invention belongs to the field of computational neuroscience, and particularly relates to a phase synchronization index-based method which comprises the following steps: a method for analyzing a network of depression and bipolar disorder brain using a phase delay index, a phase lock value and a weighted phase delay index. And constructing a functional connection matrix of the three indexes under alpha, beta, theta and delta frequency bands, and further researching and analyzing differences of different types of tested functional connections and brain networks.
Background
With the acceleration of life rhythm, the social competitiveness is getting bigger and bigger, and people have nervous and anxious moods, so that more and more patients with depressive disorder are caused, and the depressive disorder is mainly characterized clinically by remarkable and lasting mood depression and is accompanied by cognitive and behavioral damages of different degrees and higher suicide rate. While bipolar disorder patients fluctuate between mood swings (mania) and depression. The two diseases not only cause great harm to the body and the mind of patients, but also cause certain pressure to family members and the society, and simultaneously influence the good-order development of the society. Because the clinical manifestations of the disease are crossed, misdiagnosis and mistreatment are easily caused, so that the correct identification of the two diseases is of great importance.
The human brain can be regarded as a complex network, and different functional areas interact and coordinate with each other to jointly complete various complex activities of human beings. There are three main forms of network connectivity of the brain: (1) structural connection: consisting of electrical or chemical connections between neuronal synapses, such networks typically being determined using magnetic resonance data; (2) functional connection: the relevance and the statistical dependence relationship of the neural activity of a spatially separated neural unit on time are referred to; (3) effect connection: the causal effect of one nerve unit on another nerve unit is characterized, namely the two nerve units are in a regulation and regulated relation.
The measurement method of the functional brain network mainly comprises two types of linearity and nonlinearity. Because electroencephalogram signals are non-stationary nonlinear signals, a large amount of nonlinear researches are analyzed from the phase synchronization angle at present. However, the conventional Phase delay Index (PLI) is easily affected by slight fluctuation to generate signal inversion, the Phase Locking Values (PLV) are easily affected by volume conduction effect, and the Weighted Phase delay Index (WPLI) has relatively low sensitivity to real change of Phase synchronization. Because of the above disadvantages of the phase synchronization index, a feature fusion method based on three indexes is proposed herein to distinguish brain network connections of patients with depression and bipolar disorder.
Disclosure of Invention
The invention aims to provide a depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion, aiming at solving the problem of high misdiagnosis rate of depression and bipolar disorder. The provided feature fusion method achieves higher classification accuracy among healthy controls, depression and bipolar disorder patients.
The main idea for realizing the method is as follows: step one, collecting resting scalp EEG signals of an experimental group and a normal control group, wherein m electrode signals are collected for each tested group; step two, preprocessing the acquired electroencephalogram signal data, comprising the following steps: electrode localization, re-referencing, baseline removal, downsampling, band-pass filtering, ICA decomposition, and artifact removal; dividing the preprocessed signal into B frequenciesA band, and dividing the signal of each frequency band into a plurality of time segments; and step three, calculating three characteristics between the electrode signals of any i channel and any j channel in each tested electrode, each frequency band and the same time segment. In the t time slice, the phase delay index between the electrode signals of the i channel and the j channel is PLI tij Weighted phase delay index WPLI tij And a phase lock value of PLV tij (ii) a Step four: connecting three functions to a matrix R 1 ，R 2 ，R 3 Performing fusion to obtain three fused characteristics, namely PLV _ PLI, PLV _ WPLI and PLV _ PLI _ WPLI; wherein PLV _ PLI = { R = 3 ,R 1 }，PLV_WPLI＝{R 3 ,R 2 }，PLV_PLI_WPLI＝{R 3 ,R 1 ,R 2 }; step five: and sending the matrix formed by the fused features into a classifier to obtain a final classification result.
Specifically, the specific method of step two as described above is: preprocessing the acquired original brain electrical signals, comprising: electrode localization, re-referencing, baseline removal, down-sampling, band-pass filtering, ICA decomposition, and artifact removal. Dividing the frequency band of the preprocessed signal into alpha, beta, theta and delta frequency bands, wherein delta (0.5-4 Hz), theta (4-8 Hz), alpha (8-13 Hz) and beta (14-30 Hz); then, it is divided into several pieces in units of 2s.
Specifically, the phase lock value in step three described above is calculated as follows:
the method comprises the following steps: a frequency band to be tested is selected, and the EEG signals acquired by the electrodes in the same time segment are first transformed from the time domain x (t) to the frequency domain using a Hilbert transform
Step two: extracting frequency domain of each electrode of the same time segment after transformation
The obtained phase is used as the instantaneous phase of the EEG signal acquired by each electrode;
step three: calculating the instantaneous phase difference between every two electrodes of the same time segment according to the instantaneous phase obtained in the step two;
step four: calculating the phase lock value PLV between EEG signals collected by any electrodes i and j in the same time segment by using the instantaneous phase difference obtained in the third step ij 。
Specifically, the phase delay index in step three is calculated as follows:
the method comprises the following steps: a frequency band to be tested is selected, and the EEG signals acquired by the electrodes in the same time segment are first transformed from the time domain x (t) to the frequency domain using a Hilbert transform
Step two: extracting frequency domain of each electrode of the same time segment after transformation
The obtained phase is used as the instantaneous phase of the EEG signal acquired by each electrode;
step three: calculating the instantaneous phase difference between every two electrodes of the same time segment according to the instantaneous phase obtained in the step two;
step four: calculating phase delay index PLI between EEG signals collected by any electrodes i and j in the same time segment by using the instantaneous phase difference obtained in the third step ij
Specifically, the weighted phase delay index in step three described above is calculated as follows:
the method comprises the following steps: selecting discrete real-valued signals d of any two electrodes of the same segment of a tested frequency band 1 (n) and d 2 (n) to d 1 (n) and d 2 (n) performing Hilbert transform to obtain plural psi 1 (n) and psi 2 (n)；
Step two: solving for psi 1 (n) and ψ 2 (n) as P (n) + Q (n), where P (n) represents the real part and Q (n) the imaginary part;
step three: calculating the test using the imaginary part Q (n)Weighted phase delay index WPLI of any two electrodes i, j of the same time slice in one frequency band ij 。
Compared with the prior art, the method has the following advantages:
the characteristics of the electroencephalogram signals are respectively extracted by utilizing three indexes with synchronous phases and are fused according to the sample dimensions, and the fused characteristics weaken the disadvantages of each characteristic although the data volume is large, so that the typical difference of depression and bipolar disorder can be effectively identified.
Drawings
In order to more clearly explain the technical solution of the present invention, the drawings used in the present invention will be briefly described below. It is clear that these figures are only a part of the present invention.
FIG. 1 is a general flow chart of a method according to the present invention;
FIG. 2 is a graph of electrode distribution during the acquisition of resting EEG signals
FIG. 3 is a detailed diagram of the calculation process of the phase delay index, the weighted phase delay index and the phase lock value of any segment of a tested frequency band
FIG. 4R 1 Schematic view of the structure of
FIG. 5 feature fusion scheme
FIG. 6 PLI functional connectivity matrix and corresponding brain visualization network for depression, health and bipolar disorder tested in the theta band
Figure 7 WPLI functional connectivity matrix and corresponding brain visualization network for depression, health and bipolar disorder tested in the theta band
FIG. 8 PLV functional connectivity matrix and corresponding brain visualization network for depression, health and bipolar disorder tested in the theta band
Detailed Description
The present invention will now be described in detail with reference to the drawings for purposes of illustration and description. The invention is explained in detail below with reference to the figures and examples.
The invention discloses a method based on dual-channel phase synchronization feature fusionThe combined depression and bipolar disorder brain network analysis method mainly comprises the following steps: step 1, collecting resting scalp EEG signals of an experimental group and a normal control group, wherein m electrode signals are collected for each tested group; step 2, preprocessing the acquired electroencephalogram signal data, comprising the following steps: electrode localization, re-referencing, baseline removal, downsampling, band-pass filtering, ICA decomposition, and artifact removal; dividing the preprocessed signals into B frequency bands, and dividing the signals of each frequency band into a plurality of time segments; and 3, calculating three characteristics between the electrode signals of any i channel and any j channel in each tested electrode, each frequency band and the same time segment. In the t-th time segment, the phase delay index between the electrode signals of the i and j channels is PLI tij Weighted phase delay index WPLI tij And a phase lock value of PLV tij (ii) a And 4, step 4: connecting three functions to a matrix R 1 ，R 2 ，R 3 Performing fusion to obtain three fused characteristics, namely PLV _ PLI, PLV _ WPLI and PLV _ PLI _ WPLI; wherein PLV _ PLI = { R = 3 ,R 1 }，PLV_WPLI＝{R 3 ,R 2 }，PLV_PLI_WPLI＝{R 3 ,R 1 ,R 2 }; and 5: and sending the matrix formed by the fused features into a classifier to obtain a final classification result.
Step 1: and collecting data.
The data collected in this example included 35 depression patients diagnosed by a psychiatric expert, 21 bipolar disorder patients and 35 healthy age-matched controls thereof. A standard international 10-20 acquisition system is adopted to acquire 64-lead resting scalp electroencephalogram signals. During data acquisition, the electrode positions are shown in fig. 2, and the subject is always tested in a relaxed state. The sampling frequency was 5000Hz and the sampling time was about 200s.
And 2, preprocessing.
All electroencephalogram signals with a sampling rate of 5000Hz are preprocessed by using an EEGLAB tool box of a Matlab plug-in. The method comprises the following steps: electrode localization, re-referencing, baseline removal, down-sampling, band-pass filtering, ICA decomposition, and artifact removal. Wherein the band-pass filtering range is 0.5-30Hz, the down-sampling is carried out on the signals, and the sampling frequency of the processed signals is reduced to 100Hz. And dividing the signal obtained after the pretreatment into a plurality of fragments by taking 2s as a unit. In order to reduce the unstable factors of the tested signals before and after the beginning and the end of the acquisition in the experimental process, the initial 20s and the 20s before the end of the acquired electroencephalogram signals are removed, and the middle electroencephalogram signals are used for analysis and experiment.
And step 3: and calculating the phase delay index and the phase lock value of the electroencephalogram signal, and visualizing the constructed functional connection matrix and the brain network connection state.
Taking the first segment of a certain tested frequency band alpha as an example, a functional connection matrix is solved. The specific operation process is shown in fig. 3, and the following analysis is performed by combining specific examples: suppose the time domain signals of electrode 1 and electrode 2 are x respectively 1 (t)、x 2 (t), the discrete real-valued signal data lengths are all N (N = 200), and are respectively denoted by d 1 (n) and d 2 (n){n＝1,2…N}。
Step 3.1: referring first to the following formula (1), x 1 (t)、x 2 (t) substituting, namely completing the transformation of the time domain information of the first segment of the electrode 1 and the electrode 2 to the frequency domain information by using Hilbert transform;
then, the instantaneous phase of the segments is solved by using a formula (2), and the frequency domain information of the electrode 1 and the electrode 2 is substituted into the formula (2) to obtain the phase-locked loop
And &>
Then solving the instantaneous phase difference of the electrode 1 and the electrode 2 by using a formula (3);
and finally, solving the phase delay index and the phase lock value of the two segments respectively by using the formulas (4) and (5), wherein the formulas (1) to (6) involved in the method are general formulas.
Step 3.2: a weighted phase delay index is calculated. Let two time domain signals be x (t) and y (t), respectively, where the discrete real-valued signal of the same segment is denoted d 1 (n) and d 2 (N) { N =1,2 \ 8230, N, N =200}. First, the time domain signal is Hilbert transformed to obtain a complex psi by using the above formula (1) 1 (n) and psi 2 (n); then calculating the cross spectrum of the two to obtain P (n) + Q (n); and then the WPLI can be obtained by solving the formula (6).
Step 3.3: the phase delay index, the phase lock value and the weighted phase delay index between the two electrodes of the first segment have been obtained by step 3.1 and step 3.2. The calculation process of the three characteristics among the electrodes is the same, and the description is omitted here. The functional connectivity matrix R (functional connectivity matrix) shown below is obtained by the above-described procedure. The matrix is a symmetric matrix, wherein the magnitude of each element value in the matrix represents the strength of phase synchronization between two signals of the segment. The matrix form formed is shown in the following formula (7):
in the functional connection matrix, the element r pq Representing the degree of phase synchronization of the p-th electrode and the q-th electrode. We use the electrodes of the EEG signal as nodes of the brain network, r pq As a connection weight between the nodes.
Step 3.4: since the functional connectivity matrix formed by depression, bipolar disorder and healthy controls was difficult to distinguish visually, we thresholded it. Setting the PLV value to be more than 0.8 to be 1, and setting the WPLI value to be more than 0.35 to be 1 represents that information interaction exists between two electrode signals and that connection exists between nodes in the network. Conversely, setting it to 0 indicates that there is no information communication between the two electrode signals.
Step 3.5: brain connections were visualized using the brain net viewer software. Surface file uses the normalized template Brain Mesh icbm152.Nv, with Surface opacity set to 0.3 and edge opacity set to 0.7. The detailed correspondence between the electrodes and the brain regions is shown in table 1 below. The node size represents the degree of a node, namely the number of nodes directly connected with the node, and the larger the node degree is, the more important the position of the node in the network is. The different colors of the nodes represent electrodes of different brain regions. The brain is visualized using the matrix after thresholding and the differences in brain network connectivity between the experimental and control groups are analyzed to help find diseased brain regions in patients with depression and bipolar disorder.
TABLE 1 correspondence of electrodes to brain regions
And 4, step 4: the functional connection matrix formed by all the segments of the different features is converted into a matrix of 1 row m x m columns. A total of 8269 fragments were formed in 91 subjects, of which 2813 were present in depression, 2572 in healthy controls and 2884 in bipolar disorder. The size of each piece in the matrix formed by PLI and PLV is 1 x 4096, and the size of the matrix formed by WPLI is 1 x 4032 because the missing value of the matrix is deleted.
And 5: the sizes of the matrixes after feature fusion are respectively as follows: PLV _ PLI is 8269 8192, PLV _wpliis 8269 8128, PLV _pli _wpliis 8269 12224.
Step 6: classification of
And (5) sending the matrixes obtained in the steps 4 and 5 into a decision tree classifier to obtain three classification results.
To prevent overfitting, ten fold cross validation was used in the experiment, and all data were randomized into 10 groups, 9 groups for training, and 1 group for testing. Meanwhile, in order to research the stability of the model and to make the prediction result more accurate, 10 times of averaging is performed in each experimental process.
The validation of this experiment was done using data collected by the Xuanwu Hospital, and the results in Table 2 below are for the decision tree classifier at the theta band.
TABLE 2 results of experiments in the theta band
Claims (7)
1. A depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion is characterized in that: the method comprises the following steps:
the method comprises the following steps: collecting resting scalp EEG signals of an experimental group and a normal control group, wherein S tested subjects are obtained, and m electrode signals are collected in each tested subject;
step two: preprocessing the acquired electroencephalogram signal data, comprising: electrode localization, re-referencing, baseline removal, downsampling, band-pass filtering, ICA decomposition, and artifact removal; dividing the preprocessed signal into B frequency bands, dividing the signal of each frequency band into a plurality of time segments, wherein the s-th tested signal has K s A time slice;
step three: calculating three characteristics between any two channel electrode signals of i and j within each tested time segment, each frequency band and the same time segment, namely phase delay indexes, weighted phase delay indexes and phase lock values, wherein one tested time segment is tested at the t-th time segmentAnd the phase delay index between the i channel electrode signal and the j channel electrode signal is PLI tij Weighted phase delay index WPLI tij And a phase lock value of PLV tij ，t＝1、2、…、K s ；
Each frequency band of each tested corresponds to K s A phase delay index function connection matrix, K s A weighted phase delay index function connection matrix, and K s Connecting a matrix with a phase lock value function;
wherein the phase delay index function is connected with the matrix
Weighted phase delay index function connection matrix
Phase lock function connection matrix
Respectively aiming at the obtained three function connection matrixes R 1 ，R 2 ，R 3 Visualization;
step four: connecting three functions to a matrix R 1 ，R 2 ，R 3 Performing fusion to obtain fused characteristics PLV _ PLI _ WPLI = { R = 3 ,R 1 ,R 2 }；
Step five: and sending the matrix formed by the fused features into a classifier to obtain a final classification result so as to verify the model.
2. The method for analyzing the brain network of depression and bipolar disorder based on the fusion of the two-channel phase synchronization features as claimed in claim 1, wherein: the frequency band division of the second step is specifically divided into four frequency bands of delta (0.5-4 Hz), theta (4-8 Hz), alpha (8-13 Hz) and beta (14-30 Hz); the time slice is 2s.
3. The method for analyzing the brain network of depression and bipolar disorder based on the fusion of the two-channel phase synchronization features as claimed in claim 1, wherein: the specific calculation process of the phase lock value in the third step is as follows:
the method comprises the following steps: a frequency band to be tested is selected, and the EEG signals acquired by the electrodes in the same time segment are first transformed from the time domain x (t) to the frequency domain using a Hilbert transform
Step two: extracting frequency domain of each electrode transform of the same time segment
The obtained phase is used as the instantaneous phase of the EEG signal acquired by each electrode; />
Step three: calculating the instantaneous phase difference between every two electrodes of the same time segment according to the instantaneous phase obtained in the step two;
step four: calculating the phase lock value PLV between EEG signals collected by any electrodes i and j in the same time segment by using the instantaneous phase difference obtained in the third step ij 。
4. The method for analyzing the brain network of depression and bipolar disorder based on the fusion of the two-channel phase synchronization features as claimed in claim 1, wherein: the phase delay index specifically calculated in the third step is as follows:
the method comprises the following steps: a frequency band to be tested is selected, and the EEG signals acquired by the electrodes in the same time segment are first transformed from the time domain x (t) to the frequency domain using a Hilbert transform
Step two: extracting frequency domain of each electrode of the same time segment after transformation
The obtained phase is used as the instantaneous phase of the EEG signal acquired by each electrode;
step three: calculating the instantaneous phase difference between every two electrodes of the same time segment according to the instantaneous phase obtained in the step two;
step four: calculating phase delay index PLI between EEG signals collected by any electrodes i and j in the same time segment by using the instantaneous phase difference obtained in the third step ij。
5. The method for analyzing the brain network of depression and bipolar disorder based on the fusion of the two-channel phase synchronization features as claimed in claim 1, wherein: the weighted phase delay index specifically calculated in the third step is as follows:
the method comprises the following steps: selecting discrete real-valued signals d of any two electrodes of the same segment of a tested frequency band 1 (n) and d 2 (n) to d 1 (n) and d 2 (n) performing Hilbert transform to obtain complex psi 1 (n) and psi 2 (n)；
Step two: solving psi 1 (n) and psi 2 (n) as P (n) + Q (n), where P (n) represents the real part and Q (n) the imaginary part;
step three: calculating the weighted phase delay index WPLI of any two electrodes i, j of the same time segment in the tested frequency band by using the imaginary part Q (n) ij 。
6. The method for analyzing the brain network of depression and bipolar disorder based on the fusion of the two-channel phase synchronization features as claimed in claim 1, wherein: the three functional connection matrixes are visually displayed and used for visually displaying the brain network connection states of depression, bipolar disorder and health, and the three conditions are as follows:
(1) Setting a threshold, setting elements smaller than the threshold in the three functional connection matrixes as 0, and setting elements larger than the threshold as 1;
(2) Brain connectivity status was visualized using Brain Net Viewer.
7. The dual channel-based system of claim 1The method for analyzing the depression and bipolar disorder brain network with the fusion of phase synchronization features is characterized in that: the characteristics after fusion are PLV _ PLI and PLV _ WPLI, wherein PLV _ PLI = { R = { (R) 3 ,R 1 }，PLV_WPLI＝{R 3 ,R 2 }。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011633891.6A CN113017627B (en) | 2020-12-31 | 2020-12-31 | Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011633891.6A CN113017627B (en) | 2020-12-31 | 2020-12-31 | Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113017627A CN113017627A (en) | 2021-06-25 |
| CN113017627B true CN113017627B (en) | 2023-03-31 |
Family
ID=76459126
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011633891.6A Active CN113017627B (en) | 2020-12-31 | 2020-12-31 | Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113017627B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113974650A (en) * | 2021-06-29 | 2022-01-28 | 华南师范大学 | Electroencephalogram network function analysis method and device, electronic equipment and storage medium |
| CN113925517B (en) * | 2021-09-22 | 2022-08-26 | 北京脑陆科技有限公司 | Cognitive disorder recognition method, device and medium based on electroencephalogram signals |
| CN114246587A (en) * | 2021-12-22 | 2022-03-29 | 成都市第四人民医院 | Depression research and detection method |
| CN115317000A (en) * | 2022-08-31 | 2022-11-11 | 燕山大学 | Multilayer brain network model construction method based on same frequency and cross-frequency information interaction |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110013250A (en) * | 2019-04-30 | 2019-07-16 | 中南大学湘雅二医院 | A kind of multi-mode feature fusion prediction technique of depression suicide |
| CN110338785A (en) * | 2019-06-11 | 2019-10-18 | 太原理工大学 | Dynamic brain network node consistency behavior analysis method based on EEG signals |
| CN110693493A (en) * | 2019-10-12 | 2020-01-17 | 北京工业大学 | Epilepsy electroencephalogram prediction method based on convolution and recurrent neural network combined time multiscale |
| CN110859614A (en) * | 2019-11-22 | 2020-03-06 | 东南大学 | Mathematical astronomical analysis method for teenager brain network based on weighted phase lag index |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11000224B2 (en) * | 2017-09-05 | 2021-05-11 | Korea University Research And Business Foundation | Method and apparatus of monitoring anaesthesia and consciousness depth through brain network analysis |
-
2020
- 2020-12-31 CN CN202011633891.6A patent/CN113017627B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110013250A (en) * | 2019-04-30 | 2019-07-16 | 中南大学湘雅二医院 | A kind of multi-mode feature fusion prediction technique of depression suicide |
| CN110338785A (en) * | 2019-06-11 | 2019-10-18 | 太原理工大学 | Dynamic brain network node consistency behavior analysis method based on EEG signals |
| CN110693493A (en) * | 2019-10-12 | 2020-01-17 | 北京工业大学 | Epilepsy electroencephalogram prediction method based on convolution and recurrent neural network combined time multiscale |
| CN110859614A (en) * | 2019-11-22 | 2020-03-06 | 东南大学 | Mathematical astronomical analysis method for teenager brain network based on weighted phase lag index |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113017627A (en) | 2021-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113017627B (en) | Depression and bipolar disorder brain network analysis method based on two-channel phase synchronization feature fusion | |
| Ahmadlou et al. | Fuzzy synchronization likelihood-wavelet methodology for diagnosis of autism spectrum disorder | |
| CN111772585B (en) | Data extraction method for epileptic abnormal group activities in interval intracranial electroencephalogram signals | |
| Ahmed et al. | Single volume image generator and deep learning-based ASD classification | |
| Gallego-Jutglà et al. | A hybrid feature selection approach for the early diagnosis of Alzheimer’s disease | |
| CN103917159A (en) | Method and system for estimating brain concussion | |
| Deng et al. | Complexity extraction of electroencephalograms in Alzheimer's disease with weighted-permutation entropy | |
| CN114533086B (en) | Motor imagery brain electrolysis code method based on airspace characteristic time-frequency transformation | |
| Acar et al. | ACMTF for fusion of multi-modal neuroimaging data and identification of biomarkers | |
| Bugeja et al. | A novel method of EEG data acquisition, feature extraction and feature space creation for early detection of epileptic seizures | |
| Khare et al. | SchizoNET: a robust and accurate Margenau–Hill time-frequency distribution based deep neural network model for schizophrenia detection using EEG signals | |
| Gligorijević et al. | A new and fast approach towards sEMG decomposition | |
| Prasanth et al. | Deep learning for interictal epileptiform spike detection from scalp EEG frequency sub bands | |
| CN104510468A (en) | Character extraction method and device of electroencephalogram | |
| Liu et al. | A data driven approach for resting-state EEG signal classification of schizophrenia with control participants using random matrix theory | |
| CN113729729A (en) | Schizophrenia early detection system based on graph neural network and brain network | |
| Mohi-ud-Din et al. | Detection of Autism Spectrum Disorder from EEG signals using pre-trained deep convolution neural networks | |
| CN108814593B (en) | Electroencephalogram signal feature extraction method based on complex network | |
| CN116027888A (en) | P300 intention recognition method based on PLV dynamic brain function network | |
| CN113349795B (en) | Depression electroencephalogram analysis method based on sparse low-rank tensor decomposition | |
| Xu et al. | Order statistics concordance coefficient with applications to multichannel biosignal analysis | |
| Cometa et al. | Stimulus evoked causality estimation in stereo-EEG | |
| Banta et al. | A novel convolutional neural network for reconstructing surface electrocardiograms from intracardiac electrograms and vice versa | |
| Wei et al. | Automatic recognition of chewing noises in epileptic EEG based on period segmentation | |
| Shen et al. | Tensor-based EEG network formation and feature extraction for cross-session driving drowsiness detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |